Enhancing Twin Delayed Deep Deterministic Policy Gradient with Cross-Entropy Method

Hieu Trung Nguyen, Khang Tran and Ngoc Hoang Luong

Setup

• Clone this repo:

$ git clone https://github.com/junhill-2000/eTD3.git
$ cd eTD3

• The following packages are needed:

tianshou==0.4.4
gym==0.19.0
mujoco_py==2.0.2.13 (need to install mujoco200 from http://www.mujoco.org/)

Usage

train agent with eTD3

python3 mujoco_td3_mod_v3.py --task [environment_name] --seed [seed] --logdir [log_directory]

Visualize (comming soon)

Acknowledgement

Our source code is inspired by:

• CEM-RL by Pourchot et al.
• Tianshou library